Ambiguous Entity Disambiguation via Semantic and Relationship Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing natural language processing technologies face challenges in accurately disambiguating ambiguous entity words in unstructured or semi-structured data, as they struggle to simultaneously capture text semantics and entity relationships, leading to low accuracy in knowledge base disambiguation.
Innovation Solution
A text processing method and device that utilize a trained word vector model and unsupervised neural network model to generate semantic vectors and entity vectors, determining the similarity between context and candidate entities to identify the target entity, thereby improving disambiguation accuracy by learning and depicting entity information comprehensively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text overlap degree and relevance are calculated using existing knowledge base data, then disambiguation can be performed, but the accuracy is insufficient because text semantics and entity relationships cannot be simultaneously captured
Solution Approach 1:
The patent merges text semantic analysis and entity relationship analysis into a unified disambiguation framework. The semantic analysis module and entity relationship module work together, with their respective scores combined to produce the final disambiguation result, thereby simultaneously utilizing both text semantics and entity relationships for accurate disambiguation
Solution Approach 2:
The disambiguation system is segmented into independent functional modules: semantic analysis module, entity relationship module, and score integration module. This segmentation allows each module to specialize in capturing specific aspects (text semantics or entity relationships) while contributing to the overall disambiguation accuracy
2Measurement precision
If unsupervised or semi-supervised model training is used based on semantics, then disambiguation can be performed, but the complexity of the processing increases
Solution Approach 1:
The system employs unsupervised learning mechanisms where the model automatically learns entity relationships and semantic patterns from the data without requiring extensive manual annotation. The entity relationship module autonomously discovers relationships through statistical analysis, reducing the need for complex supervised training processes
Data Source
AI summary
The present disclosure provides a text processing method and device based on ambiguous entity words. The method includes: obtaining a context of a text to be disambiguated and at least two candidate entities represented by the text to be disambiguated; generating a semantic vector of the context based on a trained word vector model; generating a first entity vector of each of the at least two candidate entities based on a trained unsupervised neural network model; determining a similarity between the context and each candidate entity; and determining a target entity represented by the text to be disambiguated in the context.